As the title said. both of these things are being done manually on aws UI. Which makes it difficult if the same app needs to be replicated to other regions. Is there a conscious decision not to follow this practice?
I’m not sure of the rationale. However, whilst DynamoDB is relatively easy to set up using the Resources section using CloudFormation, Cognito is rather more complicated (at least for me!) maybe they didn’t want the distraction of disappearing down a deep CloudFormation rabbit hole.
We first go through the UI and then later in the guide use CDK to configure them. We used to do YAML but moved away from it because of how unmanageable they can be. You can find the older versions here:
I also expected to see dynamo and cognito in serverless.yml
@jayair, what you mean by using CDK. Can you give an example? Or explain why CDK is better? Probably I am going to figure it out later on the tutorial 
Thanks!
Yup we talk about it here What is AWS CDK? | Serverless Stack
We’ll probably have a serverless.yml version only as well.